Motivated by experience with students' psychological barriers to learning statistics, we modified and extended the Statistical Anxiety Rating Scale (STARS) to develop a contemporary and valid (face, content, criterion and construct) Danish measure of attitudes and relationship towards statistics for use with higher education students taking statistics within another discipline. Two subscales were excluded because of lack of conceptual unidimensionality or derogatory content, and single items were modified for face and content validity enhancement in the remaining subscales. Following a pilot study and main study, the resulting 26-item Danish instrument (HFS-R for "holdninger og forhold til statistik - Revideret", in English "Attitudes and Relationship to Statistics - Revised") consists of four subscales: Test and Class Anxiety (TCA), Interpretation Anxiety (IA), Fear of Asking for Help (FAH), and Worth of Statistics (WS). Each scale was analyzed using Rasch and graphical log-linear Rasch models. The FAH subscale fits the Rasch model, whereas the TCA, IA, and WS subscales each fit graphical log-linear Rasch models (GLLRMs) each with evidence of differential item functioning (DIF). One TCA item functioned differentially relative to age, one WS item relative to statistics course (first or second), and two IA items relative to statistics course and academic discipline (sociology, public health). The IA and TCA subscales were well targeted to the study population, while targeting of FAH and WS was poorer. Unidimensionality across the three anxiety subscales (FAH, TCA, and IA) was tested and clearly rejected. The HFS-R was found to be of sufficient psychometric quality to warrant its use in higher education research and teaching. We recommend that unidimensionality should be formally tested before using composite scores across anxiety subscales in the original STARS.